Gradient welding strength control method and system for welding galvanized steel pipe for fire fighting

Abstract

The invention discloses a gradient welding strength control method and system for a welded galvanized steel pipe for fire fighting, particularly relates to the technical field of welding automation control, and aims to solve the problem of abnormal welding seam strength gradient caused by instantaneous unbalance of an electric field and a thermal field during dynamic parameter switching in the prior art. Electromagnetic distortion characteristic quantity is extracted through frequency domain energy analysis, and thermodynamic offset state quantity is deduced in combination with a thermal diffusion trend; on the basis of frequency band correlation mapping of electromagnetic and thermodynamic parameters, thermoelectric cooperative imbalance levels are divided through mutual information entropy evolution, and a self-adaptive harmonic attenuation weight tuning instruction is generated; analyzing a mismatching relation between transient disturbance of an environment magnetic field and arc voltage modulation distortion in real time, dynamically generating a magnetic field compensation factor and reversely superposing the magnetic field compensation factor to a control instruction; and in combination with the matching degree of molten pool oscillation energy and a harmonic frequency spectrum, correcting wire feeding rate calibration, reconstructing a multi-band harmonic energy proportion spectrum, outputting an anti-interference welding control signal, and realizing dynamic balance between arc stability and molten pool heat input.

Description

Technical Field

[0001] The present invention relates to the technical field of welding automation control, and more particularly to a gradient welding strength control method and system for welding galvanized steel pipes for fire protection. Background Art

[0002] In the welding process of galvanized steel pipes for fire protection, switching of multiple welding parameters (such as current, voltage, and wire feed speed) is a common engineering practice, which aims to meet the differentiated requirements of weld strength in different areas (such as high-pressure areas and low-pressure areas); welding equipment usually performs welding tasks based on preset static parameters, or adopts simple timing switching logic when switching parameters. However, when the welding process requires dynamic adjustment of parameters in the gradient strength range (such as transitioning from a high penetration zone to a low penetration zone), the real-time response capability and parameter coordination of the welding equipment are easily affected by external disturbances.

[0003] Currently, during the dynamic switching of welding parameters, the instantaneous imbalance between the electric field (arc stability) and the thermal field (heat input to the molten pool) often leads to abnormal penetration depth or fluctuations in metallurgical bonding strength in local areas of the weld, causing the weld strength gradient to fail to meet the preset design requirements, thereby affecting the structural reliability of the fire protection pipeline and posing a safety hazard. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a gradient welding strength control method and system for welded galvanized steel pipes for fire protection to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for controlling the gradient welding strength of a welded galvanized steel pipe for fire protection comprises the following steps:

[0007] S1. Synchronously obtain the dynamic characteristics of welding current and arc light spatial distribution parameters, extract electromagnetic distortion characteristics through frequency domain energy analysis, and derive thermodynamic offset state quantities based on thermal diffusion trends;

[0008] S2. Establish a nonlinear frequency band correlation mapping between the electromagnetic distortion characteristic and the thermodynamic offset state, calculate the mutual information entropy value, and classify the thermoelectric synergy imbalance level according to the evolution of the mutual information entropy value;

[0009] S3. generating a tuning instruction set including adaptive harmonic attenuation weights according to the thermal power coordination imbalance level;

[0010] S4. Analyze the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage. When the phase amplitude of the ambient magnetic field disturbance mismatches the attenuation gradient of the arc light intensity, generate a magnetic field compensation factor.

[0011] S5. Inversely superimpose the magnetic field compensation factor on the tuning instruction set, and dynamically correct the wire feed rate calibration based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution;

[0012] S6. Reconstruct the harmonic energy ratio spectrum of the tuning instruction set and output an anti-interference welding control signal.

[0013] In a preferred embodiment, the dynamic characteristics of the welding current and the arc light spatial distribution parameters are obtained simultaneously, the electromagnetic distortion characteristics are extracted through frequency domain energy analysis, and the thermodynamic offset state quantity is derived in combination with the heat diffusion trend, including:

[0014] The welding current waveform is sampled in the time domain transient state, and the sampled data is subjected to noise reduction processing through a low-pass filter to obtain a current dynamic feature set including the instantaneous fluctuation amplitude;

[0015] The arc light radiation spatial distribution parameters are collected by a multispectral image sensor, and the time stamp alignment mechanism of the current waveform and the arc light image is used to establish a spatiotemporal synchronous mapping relationship between the current dynamic feature set and the arc light radiation spatial distribution parameters.

[0016] The fast Fourier transform is used to decompose the energy spectrum of the current dynamic feature set, extract the frequency band energy proportion from the fundamental frequency component to the fifth harmonic component, and generate the electromagnetic distortion characteristic of the arc;

[0017] Based on the spatial distribution parameters of arc radiation and the mapping relationship between the molten pool thermal radiation attenuation coefficient and the heat flux, the heat conduction rate of the welding pool along the weld axis is calculated. The heat conduction rate is compared with the pre-calibrated material thermal diffusion threshold, and the thermodynamic offset state quantity is output.

[0018] In a preferred embodiment, a nonlinear frequency band correlation mapping between the electromagnetic distortion characteristic and the thermodynamic offset state is established, the mutual information entropy value is calculated, and the thermoelectric synergy imbalance level is divided according to the evolution of the mutual information entropy value, including:

[0019] Based on the electromagnetic distortion characteristic and the thermodynamic offset state, a nonlinear frequency band correlation mapping is established, with the total energy ratio of the fundamental frequency to the fifth harmonic in the electromagnetic distortion characteristic as the input variable and the axial heat conduction rate of the welding pool in the thermodynamic offset state as the output variable.

[0020] Based on the time series of arc light radiation spatial distribution parameters after dynamic compensation, the mutual information entropy value corresponding to the frequency band correlation map is calculated using a sliding window algorithm.

[0021] According to the continuous change trend of the mutual information entropy value within the preset continuous welding time, the coupling degree level of the dynamic distribution of fundamental frequency harmonic energy and the coordination of the heat conduction rate of the molten pool during the welding process is divided. The state where the coupling degree level is lower than the preset critical level range is marked as a thermoelectric synergy imbalance state.

[0022] In a preferred embodiment, generating a tuning instruction set including adaptive harmonic attenuation weights according to the thermal power coordination imbalance level includes:

[0023] A frequency band decoupling model that matches the level of thermal power synergy imbalance is selected. The input of the frequency band decoupling model is the time series of the arc light radiation spatial distribution parameters after dynamic compensation and the real-time calculated mutual information entropy gradient value. The model outputs the initial coefficient of the harmonic attenuation weight of each frequency band.

[0024] Based on the initial coefficient of harmonic attenuation weight, the harmonic phase offset recorded in the welding current dynamic feature set is combined to perform segmented weight dynamic allocation, where the allocation amplitude value of the high-frequency harmonic segment is inversely proportional to the value of the thermoelectric synergy imbalance level;

[0025] The dynamically assigned weight coefficient is compensated and matched with the change trend of the mutual information entropy monitored in real time. The correction amount is calculated by inversely mapping the frequency band correlation between the total energy ratio from the fundamental frequency to the fifth harmonic and the axial heat conduction rate of the welding pool.

[0026] A tuning instruction set is generated according to the corrected weight coefficient, and the frequency band priority arrangement of the tuning instruction set is dynamically adjusted according to the spatial correlation strength between the axial component of the welding pool temperature gradient and the harmonic energy distribution characteristics of each frequency band.

[0027] In a preferred embodiment, the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage are analyzed. When the phase amplitude of the ambient magnetic field disturbance and the attenuation gradient of the arc light intensity are mismatched, a magnetic field compensation factor is generated, including:

[0028] Real-time collection of environmental magnetic field transient disturbance parameters and arc voltage full-band modulation distortion parameters;

[0029] The phase amplitude difference value of the environmental magnetic field disturbance is calculated through the time series data of the three-axis magnetic field intensity component of the magnetic flux sensor;

[0030] The mismatch condition is determined based on the dynamic matching degree between the time window mean value of the arc light intensity attenuation gradient parameter and the phase amplitude difference value of the environmental magnetic field disturbance;

[0031] The current-magnetic field coupling model is used to dynamically fit the magnetic field compensation factor for the spatial distribution of harmonic energy in full-band modulation distortion.

[0032] The dynamically fitted magnetic field compensation factor is injected into the weight update channel of the harmonic suppression module, and the activation threshold of the update channel is dynamically adjusted by the product of the instantaneous resistivity of the arc plasma and the magnetic field disturbance amplitude.

[0033] In a preferred embodiment, the input variable of the current-magnetic field coupling model is the convolution result of the harmonic order within a preset frequency band and the arc light radiation energy attenuation rate.

[0034] In a preferred embodiment, the magnetic field compensation factor is reversely superimposed on the tuning instruction set, and the wire feed rate calibration is dynamically corrected based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution, including:

[0035] The magnetic field compensation factor is reversely superimposed on the tuning instruction set generated by the harmonic suppression module according to the time delay parameter of the modulation waveform. The reverse superposition is achieved by deconvolution of the amplitude-frequency response curve of the compensation factor and the phase spectrum of the tuning instruction.

[0036] Synchronously collect the low-frequency fluctuation component of the molten pool oscillation energy gradient and extract the list of oscillation frequency points with the dominant proportion;

[0037] Calculate the spectrum matching degree of the harmonic distribution. The spectrum matching degree is the weighted correlation coefficient between the residual distortion amplitude after harmonic suppression and the energy of the oscillation frequency point of the molten pool.

[0038] Dynamically adjust the correction step size of wire feed rate calibration according to the amplitude value of the spectrum matching degree deviating from the preset range;

[0039] The corrected wire feeding rate calibration value is input into the feedback control loop of the wire feeding motor, and the integral time constant of the feedback control loop is adaptively configured by the inverse of the instantaneous arc length.

[0040] In a preferred embodiment, reconstructing the harmonic energy ratio spectrum of the tuning instruction set and outputting the anti-interference welding control signal includes:

[0041] Based on the dynamically corrected tuning instruction set, the energy proportion spectrum within the seventh harmonic is reconstructed based on the energy proportion of each frequency point in the molten pool oscillation frequency point list;

[0042] The tuning parameters corresponding to the harmonic frequency bands whose energy ratio difference exceeds the preset threshold are increased by the current feedback correction amount;

[0043] The electromagnetic interference suppression coefficient is superimposed on the high frequency band of the energy proportion spectrum line. The electromagnetic interference suppression coefficient is calculated based on the third-order derivative of the fluctuation amplitude of the magnetic field intensity in the welding environment.

[0044] Based on the reconstructed spectrum, a multi-band synchronously modulated anti-interference welding control signal is generated. The pulse width of the control signal is negatively correlated with the harmonic order.

[0045] Before the anti-interference control signal is output to the welding power source, a time-frequency domain cross-check is performed, and the check error is dynamically compensated by the product of the corrected wire feed rate calibration value and the instantaneous arc length.

[0046] In a preferred embodiment, the current feedback correction amount is a weighted average value of the difference between the residual distortion amplitude output by the current harmonic suppression module and the arc intensity attenuation rate in the previous sampling period.

[0047] In another aspect, the present invention provides a gradient welding strength control system for welded galvanized steel pipes for fire protection, comprising the following modules:

[0048] Feature synchronization module: used to synchronously obtain the dynamic characteristics of welding current and arc light spatial distribution parameters, extract electromagnetic distortion characteristics through frequency domain energy analysis, and derive thermodynamic offset state quantities based on thermal diffusion trends;

[0049] Entropy level mapping module: used to establish nonlinear frequency band correlation mapping between electromagnetic distortion characteristic quantities and thermodynamic offset state quantities, calculate mutual information entropy values, and classify thermoelectric synergy imbalance levels based on the evolution of mutual information entropy values;

[0050] Harmony generation module: used to generate a tuning instruction set with adaptive harmonic attenuation weights according to the thermal power coordination imbalance level;

[0051] Magnetic field compensation module: used to analyze the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage. When the phase amplitude of the ambient magnetic field disturbance mismatches the attenuation gradient of the arc light intensity, a magnetic field compensation factor is generated.

[0052] Calibration and adjustment module: used to reversely superimpose the magnetic field compensation factor into the tuning instruction set, and dynamically correct the wire feed rate calibration based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution;

[0053] Spectrum control output module: used to reconstruct the harmonic energy ratio spectrum of the tuning instruction set and output anti-interference welding control signals.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. A highly responsive closed-loop compensation system is constructed through real-time coordinated control of electromagnetic and thermodynamic characteristics, which significantly improves the control accuracy and consistency of the weld gradient intensity. By synchronously capturing the spatiotemporal correlation between the dynamic characteristics of the current and the spatial distribution of the arc light, a nonlinear mapping network of electromagnetic distortion and thermodynamic offset state quantities is constructed, which enables rapid state perception of the electric and thermal fields during welding. The thermoelectric synergistic imbalance levels are divided based on the law of mutual information entropy change, and adaptive tuning instructions with frequency band decoupling are generated, which can suppress the impact of multi-modal harmonic interference on arc stability and ensure dynamic and precise matching of the molten pool heat input gradient and the current waveform when switching parameters in different intensity sections, thereby maintaining the consistency of penetration depth and metallurgical bonding under variable working conditions.

[0056] 2. Through the reverse superposition of compensation factors and the dynamic matching of the molten pool oscillation spectrum, the coupling effect of external disturbance and internal distortion is decomposed into orthogonal control instructions in the frequency modulation domain. The unique harmonic energy ratio spectrum reconstruction technology uses multi-band anti-interference signal modulation to ensure the transmission of the main welding energy while eliminating the risk of thermal damage to the galvanized layer by higher harmonics of the third order and above. The wire feed rate calibration can dynamically correct the penetration response characteristics of the galvanized layer. Under the premise of maintaining the synchronization of rapid switching of welding parameters, the microstructure of the weld strength gradient transition area is homogenized, effectively improving the pressure resistance and sealing performance of the fire protection pipe ring weld structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a flow chart of the gradient welding strength control method of the fire-fighting welded galvanized steel pipe of the present invention;

[0058] Figure 2 This is a schematic structural diagram of the gradient welding strength control system of the welded galvanized steel pipe for fire protection according to the present invention. DETAILED DESCRIPTION

[0059] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0060] Example 1: Figure 1 A method for controlling the gradient welding strength of a welded galvanized steel pipe for fire protection according to the present invention is provided, comprising the following steps:

[0061] S1. Synchronously obtain the dynamic characteristics of welding current and arc light spatial distribution parameters, extract electromagnetic distortion characteristics through frequency domain energy analysis, and derive thermodynamic offset state quantities based on thermal diffusion trends;

[0062] S2. Establish a nonlinear frequency band correlation mapping between the electromagnetic distortion characteristic and the thermodynamic offset state, calculate the mutual information entropy value, and classify the thermoelectric synergy imbalance level according to the evolution of the mutual information entropy value;

[0063] S3. generating a tuning instruction set including adaptive harmonic attenuation weights according to the thermal power coordination imbalance level;

[0064] S4. Analyze the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage. When the phase amplitude of the ambient magnetic field disturbance mismatches the attenuation gradient of the arc light intensity, generate a magnetic field compensation factor.

[0065] S5. Inversely superimpose the magnetic field compensation factor on the tuning instruction set, and dynamically correct the wire feed rate calibration based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution;

[0066] S6. Reconstruct the harmonic energy ratio spectrum of the tuning instruction set and output an anti-interference welding control signal.

[0067] S1. Synchronously obtain the dynamic characteristics of the welding current and the arc light spatial distribution parameters, extract the electromagnetic distortion characteristics through frequency domain energy analysis, and derive the thermodynamic offset state quantity based on the heat diffusion trend. The specific implementation is as follows:

[0068] When sampling the welding current waveform in the time domain, a digital current sensor with a sampling frequency range of 10kHz to 50kHz is used to capture the transient characteristics of the dynamic changes in the current waveform. Noise reduction is performed on the raw sampled data using a low-pass filter with a cutoff frequency set to ten times the fundamental frequency of the current; for example, when the standard output frequency of the welding power supply is 50Hz, the cutoff frequency is configured to 500Hz to filter out high-frequency noise interference. The filtered current waveform is converted into a set of current dynamic characteristics containing the instantaneous fluctuation amplitude. This set is stored in a time series as structured data containing timestamps, current amplitudes, and sampling point numbers.

[0069] Arc radiation data is collected using a multispectral image sensor configured for visible and near-infrared bands. Its frame rate is synchronized with the sampling rate of the current sampling device. The imaging range of the multispectral image sensor is set to 400-700nm in the visible light band and 700-1100nm in the near-infrared band. A grayscale calibration algorithm is used to convert the light intensity signals in each band into normalized arc radiation spatial distribution parameters. During the conversion process, the multispectral image sensor collects ambient background light intensity distribution data in a non-arc state before each welding operation begins. During welding, the corresponding background value is deducted pixel by pixel from the raw light intensity data collected in real time to generate dynamically compensated arc radiation spatial distribution parameters.

[0070] The synchronous mapping of the current dynamic feature set and the spatial distribution parameters of arc light radiation is achieved through a timestamp alignment mechanism. The specific method is as follows: the IEEE 1588 precision time protocol commonly used in the industrial automation field is used to provide a unified clock source for the current sampling device and the multispectral image sensor. A synchronization timestamp tag is attached during data transmission to ensure that the synchronization error of the two data channels is controlled within ±50 microseconds. The mapped data set is stored in a multidimensional matrix containing timestamps, current amplitudes, arc light pixel coordinates, and light intensity values.

[0071] When performing frequency-domain energy spectrum decomposition on the current dynamic feature set, a fast Fourier transform (FFT) algorithm with a fixed window length of 1000 sampling points is used to convert the time-domain current signal into a frequency-domain energy spectrum. In frequency-domain analysis, the fundamental frequency component is defined as the standard output frequency of the welding power supply, such as 50 Hz or 60 Hz; the fifth harmonic component corresponds to a frequency range of 1 to 5 times the fundamental frequency, such as 50-250 Hz or 60-300 Hz. The frequency band energy percentage is calculated as follows: In the frequency-domain spectrum generated by the FFT, the amplitudes of all components within the frequency range from the fundamental frequency to the fifth harmonic are extracted. The squares of the amplitudes of each component are added and divided by the sum of the squares of the amplitudes across the entire frequency range (0 Hz to the Nyquist frequency) to obtain the total energy percentage for that interval. When the total energy proportion of the fundamental frequency to the fifth harmonic components exceeds the preset 10%, it is determined that the arc electromagnetic distortion characteristic quantity exceeds the allowable range; for example, if the energy proportion of the fundamental frequency component is less than 60% of the total energy and the third harmonic energy proportion exceeds 8%, an electromagnetic distortion characteristic quantity exceeding standard signal is generated.

[0072] When calculating thermodynamic offset state quantities based on the spatial distribution parameters of arc radiation, the material thermal diffusion threshold must be pre-calibrated. The calibration process is performed under standard welding test conditions, including: ambient temperature controlled between 20-25°C, a stable welding gun travel speed of 5-10 mm / s, three repeated welding tests using the same batch of welding materials, and the arithmetic mean of the molten pool heat conduction rate during the stable phase as the baseline threshold. During the welding process, the heat conduction rate of the welding pool along the weld axis is calculated using the spatial distribution parameters of arc radiation. The specific method is as follows: a lookup table mapping relationship is established between the molten pool heat radiation attenuation coefficient and the light intensity gradient. The corresponding heat radiation attenuation coefficient is queried based on the real-time light intensity gradient change rate, and the heat conduction rate is then deduced using the discretized Fourier heat conduction equation. For example, when the light intensity gradient in a pixel area decreases by more than 5% per second, the corresponding heat radiation attenuation coefficient increases by 0.2, and the calculated heat conduction rate is 4.0 mm per second. The rate is compared with the pre-calibrated threshold. If the deviation exceeds ±15% (for example, the allowable range is 2.98-4.03 mm when the pre-calibrated threshold is 3.5 mm per second), the thermodynamic offset state quantity is output and the process parameter adjustment instruction is triggered.

[0073] The threshold settings of all characteristic parameters are determined based on standardized test procedures, and operators are allowed to make dynamic adjustments of ±10% based on the baseline value while the equipment is running.

[0074] S2. Establish a nonlinear frequency band correlation mapping between the electromagnetic distortion characteristic and the thermodynamic offset state, calculate the mutual information entropy value, and classify the thermoelectric synergy imbalance level according to the evolution of the mutual information entropy value. The specific implementation is as follows:

[0075] In establishing a nonlinear frequency-band correlation mapping between the electromagnetic distortion characteristic and the thermodynamic offset state, based on the electromagnetic distortion characteristic and thermodynamic offset state obtained in step S1, the total energy percentage from the fundamental frequency to the fifth harmonic in the electromagnetic distortion characteristic is used as the input variable, and the axial heat conduction rate of the weld pool in the thermodynamic offset state is used as the output variable. A data-driven approach is used to establish a nonlinear frequency-band correlation mapping between the two. The total energy percentage from the fundamental frequency to the fifth harmonic of the electromagnetic distortion characteristic is extracted by fast Fourier transforming the welding current time domain signal recorded in step S1, and the frequency domain energy accumulation value is calculated. For example, if the fundamental frequency energy percentage is 50% and the fifth harmonic energy percentage is 8%, the total energy percentage is 58%.

[0076] The axial heat transfer rate of the weld pool is calculated using the melt pool temperature gradient captured by the thermal imaging sensor in step S1 and the heat conduction equation. For example, when the axial temperature gradient of the weld pool is 200°C / mm, the heat transfer rate is 1.2 mm / s. The construction of a nonlinear band-correlation mapping involves the following process: using time series data of the total energy percentage from the fundamental frequency to the fifth harmonic and the corresponding time series data of the axial heat transfer rate of the weld pool as training samples, a nonlinear regression analysis method is used to segmentally fit the relationship between the input and output variables. During the fitting process, the weight of each segmented data point is dynamically adjusted based on the variation amplitude of the arc voltage in the current dynamic feature set. For example, when the arc voltage fluctuation exceeds 5% of the reference voltage value preset in step S1 (e.g., 20V) (i.e., the change exceeds 1V), the corresponding segment weight is adjusted to 1.3 times the reference weight.

[0077] When calculating the mutual information entropy corresponding to the band-correlation map using a sliding window algorithm based on the time series of the arc radiation spatial distribution parameters after dynamic compensation, the window length and sliding step size of the sliding window are set. The window length matches the welding speed recorded in step S1. For example, when the welding speed is 5 mm / s, the window length is set to the data segment corresponding to 10 seconds. The sliding step size is consistent with the update interval of the thermodynamic offset state variable. For example, when the heat transfer rate is updated every 0.1 seconds, the window movement step size is set to 0.1 seconds.

[0078] The calculation of the mutual information entropy value includes statistical analysis of the joint discrete probability distribution of the total energy proportion from the fundamental frequency to the fifth harmonic and the axial heat conduction rate of the welding pool within the window. Specifically, the value of the total energy proportion from the fundamental frequency to the fifth harmonic is divided into 100 equal-width intervals (such as 0%-1%, 1%-2%...99%-100%), and the heat conduction rate is also divided into 100 intervals (such as 0-0.1mm / s, 0.1-0.2mm / s...9.9-10mm / s). After counting the joint frequency of the combination of the two in each window and calculating the probability value, the entropy value of the current window is accumulated according to the mutual information entropy formula. For example, the mutual information entropy value of a certain window is 0.85.

[0079] When categorizing thermoelectric synergy imbalance levels based on the continuous variation trend of the mutual information entropy over a preset continuous welding duration, a baseline fluctuation range for the mutual information entropy value is pre-determined through a standard welding test in step S1. The baseline range is calibrated by continuously collecting mutual information entropy data for 15 minutes during a normal, uninterrupted welding phase. The average value is calculated to be 1.2, the standard deviation is 0.15, and the baseline range is set to 1.2 ± 0.225 (i.e., the average value ± 1.5 times the standard deviation).

[0080] The coupling level classification rules are as follows: when the entropy variance of 20 consecutive windows is less than the baseline variance (0.0225), the coupling level is high; when the variance is between 0.0225 and 0.045, the coupling level is medium; and when the variance exceeds 0.045 and a single upward or downward trend is observed for 10 consecutive windows, the coupling level is low. When the welding stage corresponding to the low coupling level is marked as a thermoelectric synergistic imbalance state, it is necessary to simultaneously verify whether the axial heat transfer rate of the weld pool exceeds the allowable range calibrated in step S1 (for example, if the baseline rate is 1.2 mm / s, the allowable deviation is ±15%, i.e., 1.02-1.38 mm / s). The imbalance state is triggered when both conditions are met.

[0081] The preset continuous welding time corresponds to the weld length. For example, the preset time for a 500mm weld at a welding speed of 10mm / s is 50 seconds. The number of discretized statistical units in the sliding window is adjusted according to the parameter resolution. For example, when the acquisition accuracy of the total energy percentage from the fundamental frequency to the fifth harmonic is 0.1%, the number of units is set to 1000 to ensure accuracy. The dynamic weight adjustment mechanism of the band-related mapping is further linked to the arc light radiation parameters. For example, when the light intensity gradient after dynamic compensation exceeds 500lux / mm, the weight coefficient of the corresponding segment is increased in steps of 0.05 until the gradient returns to below 300lux / mm.

[0082] S3. Generate a tuning instruction set containing adaptive harmonic attenuation weights according to the thermal power coordination imbalance level, which is specifically implemented as follows:

[0083] When a tuning instruction set including adaptive harmonic attenuation weights is generated according to the currently detected thermoelectric coordination imbalance level, the following implementation process is specifically included. Based on the numerical range of the thermoelectric synergy imbalance level determined in real time, an adaptive model is selected from multiple pre-calibrated frequency band decoupling models, where different models correspond to different levels of imbalance. The input data of the frequency band decoupling model includes a time series of the spatial distribution parameters of the arc light radiation after dynamic compensation. This is obtained by collecting the arc light radiation intensity waveform through a multispectral sensor array arranged around the welding device, then synchronously calculating the arc light energy distribution parameters in each spatial direction according to a time window using a spatial discrete integration algorithm, and generating a time series of these parameters according to a preset period. Another input data of the frequency band decoupling model is the real-time calculated mutual information entropy gradient value, which is calculated by performing a window sliding mutual information entropy analysis on the synchronously sampled data of the welding current and arc voltage, and taking the first-order derivative of the entropy value change curve after fitting the least squares polynomial as the gradient value. The output of the frequency band decoupling model is the initial harmonic attenuation weight coefficient for each frequency band. The frequency band division corresponding to the initial harmonic attenuation weight coefficient is aligned with the standard harmonic frequency range of the welding system, including the interval from 100Hz covering the fundamental frequency to 500Hz covering the fifth harmonic.

[0084] After obtaining the initial coefficient of the harmonic attenuation weight of each frequency band, the segmented weight dynamic allocation is completed in combination with the harmonic phase offset recorded in the welding current dynamic feature set, where the welding current dynamic feature set is the amplitude and phase information including the fundamental frequency to the fifth harmonic extracted by fast Fourier transform after the welding current waveform is collected by a high-precision current sensor. The harmonic phase offset is the difference between the current harmonic phase and the expected standard phase of the welding power supply; in the process of segmented weight dynamic allocation, a rule is set that the allocation amplitude value of the high-frequency harmonic segment is inversely proportional to the value of the thermoelectric synergy imbalance level. Specifically, when the thermoelectric synergy imbalance level is low, the weight allocation amplitude value of the high-frequency segment is set to be greater than 1.5 times the fundamental frequency weight; when the thermoelectric synergy imbalance level increases to a high level, the weight allocation amplitude value of the high-frequency segment is gradually reduced to less than 0.6 times the fundamental frequency weight according to the preset coefficient, so as to achieve differentiated suppression of high-frequency harmonics.

[0085] After completing the dynamic weight allocation, the assigned weight coefficients are compensated and matched with the mutual information entropy change trend. The implementation process of compensation matching includes: obtaining the current real-time monitoring mutual information entropy change trend curve, and determining the directionality of the mutual information entropy change trend by the sign of the gradient value within the continuous time window; inversely calculating the correction value based on the frequency band correlation mapping relationship between the total energy ratio of the fundamental frequency to the fifth harmonic and the axial heat conduction rate of the weld pool. The total energy ratio of the fundamental frequency to the fifth harmonic is calculated by accumulating the square sum of the energy amplitudes of the fundamental frequency and each harmonic from the harmonic analysis results, and then calculating the proportion of the fundamental frequency in the total; the method for obtaining the axial heat conduction rate of the weld pool is: using an infrared thermal imager to capture the surface temperature field distribution of the weld pool, combining the heat conduction equation to calculate the temperature drop gradient in the axial direction, and then converting it into a conduction rate value based on the material thermophysical parameters; the frequency band correlation mapping relationship is constructed by experimentally measuring the correspondence table between different harmonic energy ratios and corresponding heat conduction rates in the initial calibration stage, and determining the specific correction value under the current ratio conditions through an interpolation algorithm during the operation stage.

[0086] When generating the tuning instruction set based on the compensated weight coefficient, the frequency band priority arrangement is dynamically adjusted based on the spatial correlation strength between the axial component of the welding pool temperature gradient and the harmonic energy distribution characteristics of each frequency band; the spatial correlation strength is calculated as follows: for the harmonic energy distribution characteristics of each frequency band, the energy contribution weight of different spatial nodes in the welding area is extracted, and the Pearson correlation coefficient is calculated with the component of the axial temperature gradient of the molten pool at each spatial node, and the absolute value of the correlation coefficient is set as the judgment value of the spatial correlation strength; the generation of the tuning instruction set includes the normalization of the weight coefficients of all frequency bands. The normalization process is to limit the weight coefficients of all frequency bands to the range of 0 to 1, and output them to the welding power supply controller after weighted sorting according to priority. The welding power supply controller executes the tuning instruction set through the inverter circuit switching frequency modulation strategy of the harmonic suppression module.

[0087] S4. Analyze the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage. When the phase amplitude of the ambient magnetic field disturbance is mismatched with the arc light intensity attenuation gradient, generate a magnetic field compensation factor. The specific implementation is as follows:

[0088] First, the specific implementation process for real-time acquisition of transient disturbance parameters of the ambient magnetic field using a three-axis magnetic flux sensor is as follows: the three probes of an industrial-grade three-axis magnetic flux sensor are arranged in orthogonal orientations on the top surface of the protective housing of the welding device, so that their sensing coordinate system is aligned with the magnetic field direction of the welding arc region. The magnetic field intensity component of each axis is sampled at a 0.01-second interval to form time series data, consisting of 6,000 data points recorded within 1 minute. Before each acquisition, the sensor's zero-point drift error is eliminated through a Hall effect calibration module. The obtained full-band modulation distortion parameters of the arc voltage are derived from the wideband spectrum analysis results of the PWM carrier signal modulation waveform output by the welding power controller in the previous step. The wideband spectrum analysis converts the time-domain voltage signal within the sampling window into a frequency-energy distribution spectrum covering the fundamental frequency to the seventh harmonic through a fast Fourier transform, and records the amplitude offset of each harmonic relative to the fundamental frequency.

[0089] Secondly, the dynamic calculation of the phase and amplitude difference of the environmental magnetic field disturbance is implemented by selecting a continuous time period of 32 sampling periods from the time series data of the three-axis magnetic field intensity components as a calculation window. Within each calculation window, the Hilbert transform algorithm is used to extract the phase time-varying curves and amplitude envelopes corresponding to the X, Y, and Z axes. The slope of the phase time-varying curve represents the rate of change of the magnetic field for each axis, and the mean of the amplitude envelope represents the average magnetic field intensity for each axis. The total phase and amplitude variation trend of the composite magnetic field disturbance is obtained by vector superposition of the change rates of the three axes. The phase and amplitude difference of the composite magnetic field disturbance at the current moment is determined by comparing the phase and amplitude differences of the composite magnetic field between two adjacent time windows. If the total phase and amplitude of the composite magnetic field continuously changes by more than a preset 20% threshold within 5 seconds, a real-time update instruction for the phase and amplitude difference is triggered.

[0090] The specific implementation method for determining the dynamic matching degree of the arc light intensity attenuation gradient parameter is as follows: the arc light intensity attenuation gradient parameter is obtained from the continuous monitoring data based on the multispectral sensor array established in the previous step, and the calculation period of its time window mean is synchronized with the update period of the environmental magnetic field disturbance phase amplitude difference value. Within the synchronization period, the dynamic matching degree is determined based on the consistency of the change direction of the two. For example, when the mean value of the arc light intensity attenuation gradient increases within a preset period, while the environmental magnetic field disturbance phase amplitude difference value decreases within the same period, the mismatch condition is determined to be met through the normalized difference comparison formula. The specific implementation form of the difference comparison formula includes: after normalizing the time series of the two parameters, calculating the covariance matrix within the sliding window at the same time node. When the main diagonal elements of the covariance matrix have opposite signs, it is marked as a mismatch state. The normalization process includes linear normalization to remove the dimension effect, for example, mapping the arc light intensity attenuation gradient parameter to the range of 0-1 and mapping the environmental magnetic field disturbance phase amplitude difference value to the symmetrical interval of -0.5 to 0.5.

[0091] The process for constructing a current-magnetic field coupling model and dynamically fitting the magnetic field compensation factor involves: The current-magnetic field coupling model is based on a pre-calibrated dataset of harmonic orders and arc radiation energy decay rates. The input variable is the convolution of these two factors within a preset frequency band, which was previously defined as the set of all integer harmonics between 100 Hz and 700 Hz. The convolution calculation is specifically implemented by multiplying the actual orders of the fundamental to seventh harmonics detected in the current welding current (e.g., when the fundamental frequency is 100 Hz, the seventh harmonic is 700 Hz) by the arc radiation energy decay rate of the same frequency (e.g., the arc radiation energy decay rate at 100 Hz is 0.2 units / second), and then accumulating the products to obtain a weighted impact factor for the specific frequency band. During the dynamic fitting process, the weight distribution ratios of different harmonic frequency bands are adjusted according to the triggering frequency of the mismatch condition. For example, when the mismatch condition is triggered more than three times within 10 seconds, the initial weight ratio of the fifth harmonic is increased from 0.3 to 0.45, and the weight ratio of the third harmonic is reduced. The dynamic parameters of the magnetic field compensation factor are updated through the iterative optimization of the least squares method.

[0092] The specific operations for injecting the magnetic field compensation factor and adjusting the activation threshold of the weight update channel are as follows: the dynamically fitted magnetic field compensation factor is converted to a 32-bit floating-point format and transmitted to the central controller of the harmonic suppression module via a digital signal interface. The activation threshold of the weight update channel is dynamically adjusted based on the product of the instantaneous resistivity of the arc plasma and the current magnetic field disturbance amplitude. The instantaneous resistivity of the arc plasma is calculated by the real-time ratio of the arc voltage and current sampling values. For example, when the current is 250A and the arc voltage is 26V, the instantaneous resistivity is 0.104 ohms. The specific configuration method for dynamically adjusting the product includes adjusting the threshold according to a preset ratio based on the offset of the product result based on the baseline value of the activation threshold (for example, set to 1.2 units). For example, when the product result increases to 1.5 times the baseline value, the corresponding activation threshold is increased to 80% of the baseline value; when the product decreases to 0.7 times the baseline value, the threshold is reduced to 65% of the baseline value. The activation judgment condition of the weight update channel is additionally equipped with a time hysteresis characteristic, that is, after the current magnetic compensation factor parameters remain stable for more than 2 seconds, the forced synchronization state of the weight update process is automatically released and the next compensation cycle is entered.

[0093] S5. The magnetic field compensation factor is reversely superimposed on the tuning instruction set. Based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution, the wire feed rate calibration is dynamically corrected. The specific implementation is as follows:

[0094] First, the implementation process of the reverse superposition of the magnetic field compensation factor is as follows: the magnetic field compensation factor is derived from the 32-bit floating-point compensation parameter dynamically fitted by the current-magnetic field coupling model in the previous step. The tuning instruction set is generated in real time by the harmonic suppression module according to the harmonic distortion characteristics of the current welding machine output current, which contains a suppression amplitude instruction sequence for the fundamental to seventh harmonics. The specific implementation method of reverse superposition includes: using the preset modulation waveform delay parameter to compensate the phase delay of the magnetic field compensation factor on the time axis. For example, when the delay parameter is 0.05 seconds, the timing data of the compensation factor is shifted back by 50 milliseconds as a whole to match the actual effective time window of the tuning instruction set; further, the frequency domain amplitude distribution characteristics of the compensation factor are reversely loaded into the tuning instruction by using the amplitude-frequency response curve and phase spectrum deconvolution operation. The deconvolution is achieved through the inverse operation of frequency domain multiplication. For example, at a frequency of 100 Hz, if the amplitude-frequency response value of the compensation factor is 1.2, and the phase spectrum component of the tuning instruction corresponding to the frequency point is 0.8 radians, then the amplitude of the tuning instruction at this frequency point after reverse superposition is adjusted to 1.2 times the original value, and the phase is adjusted to the difference between the original radian value and 0.2 radians.

[0095] The acquisition and processing of the low-frequency fluctuation components of the molten pool's oscillation energy gradient involves real-time acquisition of the molten pool's surface vibration acceleration signal via a piezoelectric vibration sensor mounted on the side of the welding table. The acceleration signal is sampled at a rate of 1000 Hz. After acquisition, the signal is filtered through a high-pass filter to remove mechanical vibration noise below 10 Hz, retaining the effective oscillation components in the 30 Hz to 200 Hz range. The dominant oscillation frequency list is extracted using a frequency-domain peak search method, which detects local maxima in the power spectrum density curve of the acceleration signal and selects frequencies with energy values ​​exceeding twice the average energy of the overall spectrum as the dominant frequencies. For example, if the energies of the 125 Hz and 180 Hz frequencies are detected to be 2.5 and 3 times the average, respectively, these two frequencies are designated as the dominant oscillation frequencies within the current time window.

[0096] The specific implementation method for calculating the spectral matching degree of harmonic distribution is as follows: the residual distortion amplitude after harmonic suppression is obtained by real-time spectral analysis of the actual waveform of arc voltage modulation, and the length of the analysis window is consistent with the processing window length of the molten pool oscillation signal to ensure time synchronization consistency; in the calculation of the weighted correlation coefficient, the distribution of weights is determined according to the energy proportion of each frequency point in the dominant frequency point list. For example, when 125Hz accounts for 60% of the total energy and 180Hz accounts for 30%, the corresponding weight coefficients are 0.6 and 0.3 respectively, and the remaining frequency points share a weight of 0.1; the calculation method of the correlation coefficient is based on the ratio of the cumulative sum of the amplitude products of the two spectral sequences at each frequency point to the geometric mean of the squares of their respective amplitudes. The calculation result is mapped to the interval of -1 to 1. A value close to 1 indicates a high spectral matching degree.

[0097] The dynamic adjustment logic for the wire feed rate calibration correction step size includes: a preset range is set to an absolute value of the correlation coefficient of no less than 0.7. When the correlation coefficient is detected to be below 0.6, the spectrum matching is determined to have deviated from the preset range and a correction mechanism is triggered. In calculating the adjustment coefficient, the arc light intensity decay rate is derived from the light intensity change rate of the multispectral sensor in the 800nm ​​band. The slope of the linear fit is calculated by sampling 10 times per second. The gradient change rate of the molten pool surface tension is obtained through thermal imaging analysis of the molten pool profile. This involves differentially calculating the change in the curvature radius of consecutive frames of the geometric shape of the molten pool edge. For example, if the curvature radius increases from 10 mm to 12 mm within 0.1 seconds, the gradient change rate is 20 mm / s. The product of the two is used to scale the baseline correction step size. For example, when the decay rate is 0.5 units / second and the gradient change rate is 15 mm / s, the adjustment coefficient is 7.5. If the baseline step size is 0.2 mm / s, the actual correction step size is adjusted to 1.5 mm / s.

[0098] The parameter configuration process for the wire feed motor feedback control loop includes: The instantaneous arc length is calculated by multiplying the arc voltage to current ratio by a preset arc characteristic coefficient. For example, when the arc voltage is 28V and the current is 300A, the instantaneous arc length is 28 / 300 × 0.12, which is a coefficient of ≈ 6.72mm. The reciprocal configuration of the integral time constant ensures that the control loop's response speed is adaptively adjusted as the arc length changes. For example, when the arc length is shortened to 5mm, the integral time constant is set to 0.2 seconds, while when the arc length increases to 8mm, the integral time constant is extended to 0.35 seconds to avoid oscillation instability in the melt pool caused by sudden changes in wire feed rate. The parameter update cycle involved in this entire process is strictly synchronized with the processing window of the melt pool oscillation signal to ensure the timing matching of the correction operation.

[0099] S6. Reconstruct the harmonic energy ratio spectrum of the tuning instruction set and output the anti-interference welding control signal, which is specifically implemented as follows:

[0100] In the process of reconstructing the energy proportion spectrum based on the dynamically corrected tuning instruction set, the dynamically corrected tuning instruction set is derived from the harmonic suppression parameter set dynamically adjusted by spectrum matching in the previous step. The corrected tuning instruction contains a sequence of harmonic suppression amplitude values ​​within the seventh order after amplitude-frequency response deconvolution processing; the energy proportion of each frequency point in the molten pool oscillation frequency point list is obtained by real-time calculation of the ratio of the energy of the dominant frequency point to the total energy after spectrum analysis of the acceleration signal collected by the vibration sensor. For example, when the 125Hz frequency point accounts for 63% of the total energy, the energy proportion spectrum is set to 63% at 125Hz, and the remaining frequency points are filled according to the actual proportion; the energy distribution ratio of each harmonic in the reconstructed spectrum corresponds to the actual energy distribution of the molten pool oscillation, forming a proportion mapping relationship spectrum from the fundamental wave to the seventh harmonic.

[0101] When the tuning parameters corresponding to the harmonic frequency bands whose energy ratio difference exceeds the preset threshold are increased by the current feedback correction amount, the preset threshold is set to a floating range of 20% to 30% of the rated energy ratio of the harmonic frequency band. For example, when the rated ratio of a third harmonic should be 15%, if its actual ratio exceeds 19.5%, the correction is triggered; in the calculation of the current feedback correction amount, the residual distortion amplitude is obtained by averaging the distortion component amplitudes of the real-time monitoring waveform spectrum output by the harmonic suppression module. For example, the third harmonic distortion in a certain period of time The amplitude is 2.1V, and the difference in arc light intensity decay rate is the difference between the light intensity change rate of the previous sampling period and the current period. If the decay rate of the previous period is 0.3 units / second and the current one is 0.25 units / second, the difference is 0.05 units / second. The weight of the weighted average value is set to 60% for the residual distortion amplitude and 40% for the light intensity difference. If the current residual distortion is 2.1V and the difference is 0.05, the correction amount is 2.1×0.6+0.05×0.4=1.28.

[0102] In the specific method of superimposing the electromagnetic interference suppression coefficient on the high-frequency band, the calculation of the electromagnetic interference suppression coefficient depends on the real-time monitoring data of the magnetic field strength of the welding environment. The magnetic field strength fluctuation amplitude is collected by the magnetic induction coil sensor at a sampling rate of 20 times per second and its absolute fluctuation amount is calculated; the calculation process of the third-order derivative is to perform three differential operations on the curve of the fluctuation amplitude changing with time. For example, when the fluctuation amplitude of three adjacent sampling points is 0.02T, 0.025T, and 0.03T, the first-order derivative is 0.005T / s and 0.005T / s, the second-order derivative is 0T / s², and the third-order derivative is 0T / s³; the generation of the suppression coefficient is scaled according to the absolute value of the derivative. If the third-order derivative is zero, the suppression coefficient remains at the preset base value, such as 0.8.

[0103] When generating multi-band synchronous modulation anti-interference control signals, the negative correlation between pulse width and harmonic order is achieved by gradually shortening the pulse width as the harmonic order increases, tuning the parameters from the fundamental to the seventh harmonic. For example, the pulse width of the fundamental is set to 200 microseconds, the second harmonic to 180 microseconds, and the seventh harmonic to 100 microseconds. Multi-band synchronous modulation is achieved through time-division multiplexing control logic, which assigns pulses of different harmonic modulations to the same carrier signal in a time-slice sequence. The time-slice length is consistent with the corresponding harmonic pulse width, ensuring that the timing of each harmonic modulation does not overlap.

[0104] When performing time-frequency domain cross-checking, the generation of the check error comes from the deviation detection between the reconstructed spectrum line and the actual output signal spectrum. The deviation is calculated by the percentage difference between the command amplitude and the actual waveform amplitude of each harmonic frequency band; in the dynamic compensation product, the corrected wire feeding rate calibration value is the wire feeding motor speed command value finally adjusted in the previous step, for example, 3.5 meters / minute, and the instantaneous arc length is the result of multiplying the arc voltage to current ratio by the arc characteristic coefficient. For example, when the arc voltage is 30V, the current is 280A and the coefficient is 0.1, the length is about 1.07 mm; if the product result is 3.5×1.07≈3.75 mm·m / (minute·unit), this value is fed back to the check error correction loop as the compensation amount, reducing the time domain deviation of the signal to within 0.3 times the original error; the synchronization mechanism of the iterative cycle is reflected in the immediate start of the compensation operation after each processing window of the molten pool oscillation signal ends, for example, error correction is performed every 0.1 seconds. Throughout the entire process, the generation, verification, and compensation of control signals all use the same time reference signal source as the previous steps to ensure timing consistency of data processing at each stage.

[0105] Example 2: Figure 2 The structural diagram of the gradient welding strength control system of the fire-fighting welded galvanized steel pipe of the present invention is given. The gradient welding strength control system of the fire-fighting welded galvanized steel pipe includes the following modules:

[0106] Feature synchronization module: used to synchronously obtain the dynamic characteristics of welding current and arc light spatial distribution parameters, extract electromagnetic distortion characteristics through frequency domain energy analysis, and derive thermodynamic offset state quantities based on thermal diffusion trends;

[0107] Entropy level mapping module: used to establish nonlinear frequency band correlation mapping between electromagnetic distortion characteristic quantities and thermodynamic offset state quantities, calculate mutual information entropy values, and classify thermoelectric synergy imbalance levels based on the evolution of mutual information entropy values;

[0108] Harmony generation module: used to generate a tuning instruction set with adaptive harmonic attenuation weights according to the thermal power coordination imbalance level;

[0109] Magnetic field compensation module: used to analyze the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage. When the phase amplitude of the ambient magnetic field disturbance mismatches the attenuation gradient of the arc light intensity, a magnetic field compensation factor is generated.

[0110] Calibration and adjustment module: used to reversely superimpose the magnetic field compensation factor into the tuning instruction set, and dynamically correct the wire feed rate calibration based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution;

[0111] Spectrum control output module: used to reconstruct the harmonic energy ratio spectrum of the tuning instruction set and output anti-interference welding control signals.

[0112] The above formulas are all dimensionless and numerical calculations. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0113] It should be noted that the present invention can be deployed on the device itself to implement embedded applications, and can also be run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0114] The above embodiments can be implemented in whole or in part via software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in the embodiments of this application are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired means (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0115] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and modules described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0116] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0117] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, and may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0118] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0119] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0120] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

[0121] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for controlling the gradient welding strength of a welded galvanized steel pipe for fire protection, characterized in that: The steps include: S1. Synchronously obtain the dynamic characteristics of welding current and arc light spatial distribution parameters, extract electromagnetic distortion characteristics through frequency domain energy analysis, and derive thermodynamic offset state quantities based on thermal diffusion trends; S2. Establish a nonlinear frequency band correlation mapping between the electromagnetic distortion characteristic and the thermodynamic offset state quantity, calculate the mutual information entropy value, and classify the thermoelectric synergy imbalance level according to the evolution of the mutual information entropy value; S3. Generate a tuning instruction set including adaptive harmonic attenuation weights according to the thermal power coordination imbalance level; S4. Analyze the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage. When the phase amplitude of the ambient magnetic field disturbance mismatches the attenuation gradient of the arc light intensity, generate a magnetic field compensation factor. S5. Inversely superimpose the magnetic field compensation factor on the tuning instruction set, and dynamically correct the wire feed rate calibration based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution; S6. Reconstruct the harmonic energy ratio spectrum of the tuning instruction set and output an anti-interference welding control signal.

2. The gradient welding strength control method of the fire-fighting welded galvanized steel pipe according to claim 1, characterized in that: The dynamic characteristics of welding current and arc light spatial distribution parameters are acquired simultaneously. The electromagnetic distortion characteristics are extracted through frequency domain energy analysis. The thermodynamic offset state is derived based on the thermal diffusion trend, including: The welding current waveform is sampled in the time domain transient state, and the sampled data is subjected to noise reduction processing through a low-pass filter to obtain a current dynamic feature set including the instantaneous fluctuation amplitude; The arc light radiation spatial distribution parameters are collected by a multispectral image sensor, and the time stamp alignment mechanism of the current waveform and the arc light image is used to establish a spatiotemporal synchronous mapping relationship between the current dynamic feature set and the arc light radiation spatial distribution parameters. The fast Fourier transform is used to decompose the energy spectrum of the current dynamic feature set, extract the frequency band energy proportion from the fundamental frequency component to the fifth harmonic component, and generate the electromagnetic distortion characteristic of the arc; Based on the spatial distribution parameters of arc radiation and the mapping relationship between the molten pool thermal radiation attenuation coefficient and the heat flux, the heat conduction rate of the welding pool along the weld axis is calculated. The heat conduction rate is compared with the pre-calibrated material thermal diffusion threshold, and the thermodynamic offset state quantity is output.

3. The gradient welding strength control method of the fire-fighting welded galvanized steel pipe according to claim 1, characterized in that: A nonlinear frequency band correlation mapping between electromagnetic distortion characteristic quantities and thermodynamic offset state quantities is established, and the mutual information entropy value is calculated. The thermoelectric synergy imbalance level is classified according to the evolution of the mutual information entropy value, including: Based on the electromagnetic distortion characteristic and the thermodynamic offset state, a nonlinear frequency band correlation mapping is established, with the total energy ratio of the fundamental frequency to the fifth harmonic in the electromagnetic distortion characteristic as the input variable and the axial heat conduction rate of the welding pool in the thermodynamic offset state as the output variable. Based on the time series of arc light radiation spatial distribution parameters after dynamic compensation, the mutual information entropy value corresponding to the frequency band correlation map is calculated using a sliding window algorithm. According to the continuous change trend of the mutual information entropy value within the preset continuous welding time, the coupling degree level of the dynamic distribution of fundamental frequency harmonic energy and the coordination of the heat conduction rate of the molten pool during the welding process is divided. The state where the coupling degree level is lower than the preset critical level range is marked as a thermoelectric synergy imbalance state.

4. The gradient welding strength control method of a fire-fighting welded galvanized steel pipe according to claim 1, characterized in that: Generates a tuning instruction set with adaptive harmonic attenuation weights based on the heat and power coordination imbalance level, including: A frequency band decoupling model that matches the level of thermal power synergy imbalance is selected. The input of the frequency band decoupling model is the time series of the arc light radiation spatial distribution parameters after dynamic compensation and the real-time calculated mutual information entropy gradient value. The model outputs the initial coefficient of the harmonic attenuation weight of each frequency band. Based on the initial coefficient of harmonic attenuation weight, the harmonic phase offset recorded in the welding current dynamic feature set is combined to perform segmented weight dynamic allocation, where the allocation amplitude value of the high-frequency harmonic segment is inversely proportional to the value of the thermoelectric synergy imbalance level; The dynamically assigned weight coefficient is compensated and matched with the change trend of the mutual information entropy monitored in real time. The correction amount is calculated by inversely mapping the frequency band correlation between the total energy ratio from the fundamental frequency to the fifth harmonic and the axial heat conduction rate of the welding pool. A tuning instruction set is generated according to the corrected weight coefficient, and the frequency band priority arrangement of the tuning instruction set is dynamically adjusted according to the spatial correlation strength between the axial component of the welding pool temperature gradient and the harmonic energy distribution characteristics of each frequency band.

5. The gradient welding strength control method of the fire-fighting welded galvanized steel pipe according to claim 1, characterized in that: Analyze the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage. When the phase amplitude of the ambient magnetic field disturbance mismatches the arc light intensity attenuation gradient, generate a magnetic field compensation factor, including: Real-time collection of transient disturbance parameters of the ambient magnetic field and full-band modulation distortion parameters of the arc voltage; The phase amplitude difference value of the environmental magnetic field disturbance is calculated through the time series data of the three-axis magnetic field intensity component of the magnetic flux sensor; The mismatch condition is determined based on the dynamic matching degree between the time window mean value of the arc light intensity attenuation gradient parameter and the phase amplitude difference value of the environmental magnetic field disturbance; The current-magnetic field coupling model is used to dynamically fit the magnetic field compensation factor for the spatial distribution of harmonic energy in full-band modulation distortion. The dynamically fitted magnetic field compensation factor is injected into the weight update channel of the harmonic suppression module, and the activation threshold of the update channel is dynamically adjusted by the product of the instantaneous resistivity of the arc plasma and the magnetic field disturbance amplitude.

6. The gradient welding strength control method of a fire-fighting welded galvanized steel pipe according to claim 5, characterized in that: The input variables of the current-magnetic field coupling model are the convolution results of the harmonic order within the preset frequency band and the arc light radiation energy attenuation rate.

7. The gradient welding strength control method of a fire-fighting welded galvanized steel pipe according to claim 1, characterized in that: The magnetic field compensation factor is reversely superimposed on the tuning instruction set, and the wire feed rate calibration is dynamically corrected based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution, including: The magnetic field compensation factor is reversely superimposed on the tuning instruction set generated by the harmonic suppression module according to the time delay parameter of the modulation waveform. The reverse superposition is achieved by deconvolution of the amplitude-frequency response curve of the compensation factor and the phase spectrum of the tuning instruction. Synchronously collect the low-frequency fluctuation component of the molten pool oscillation energy gradient and extract the list of oscillation frequency points with the dominant proportion; Calculate the spectrum matching degree of the harmonic distribution. The spectrum matching degree is the weighted correlation coefficient between the residual distortion amplitude after harmonic suppression and the energy of the oscillation frequency point of the molten pool. Dynamically adjust the correction step size of wire feed rate calibration according to the amplitude value of the spectrum matching degree deviating from the preset range; The corrected wire feeding rate calibration value is input into the feedback control loop of the wire feeding motor, and the integral time constant of the feedback control loop is adaptively configured by the inverse of the instantaneous arc length.

8. The gradient welding strength control method for welded galvanized steel pipes for fire protection according to claim 1, characterized in that: Reconstruct the harmonic energy ratio spectrum of the tuning instruction set and output the anti-interference welding control signal, including: Based on the dynamically corrected tuning instruction set, the energy proportion spectrum within the seventh harmonic is reconstructed based on the energy proportion of each frequency point in the molten pool oscillation frequency point list; The tuning parameters corresponding to the harmonic frequency bands whose energy ratio difference exceeds the preset threshold are increased by the current feedback correction amount; The electromagnetic interference suppression coefficient is superimposed on the high frequency band of the energy proportion spectrum line. The electromagnetic interference suppression coefficient is calculated based on the third-order derivative of the fluctuation amplitude of the magnetic field intensity in the welding environment. Based on the reconstructed spectrum, a multi-band synchronously modulated anti-interference welding control signal is generated. The pulse width of the control signal is negatively correlated with the harmonic order. Before the anti-interference control signal is output to the welding power source, a time-frequency domain cross-check is performed, and the check error is dynamically compensated by the product of the corrected wire feed rate calibration value and the instantaneous arc length.

9. The gradient welding strength control method of a fire-fighting welded galvanized steel pipe according to claim 8, characterized in that: The current feedback correction amount is the weighted average of the difference between the residual distortion amplitude output by the current harmonic suppression module and the arc intensity attenuation rate of the previous sampling period.

10. A gradient welding strength control system for welded galvanized steel pipes for firefighting, used to implement the gradient welding strength control method for welded galvanized steel pipes for firefighting according to any one of claims 1 to 9, characterized in that: Includes the following modules: Feature synchronization module: used to synchronously obtain the dynamic characteristics of welding current and arc light spatial distribution parameters, extract electromagnetic distortion characteristics through frequency domain energy analysis, and derive thermodynamic offset state quantities based on thermal diffusion trends; Entropy level mapping module: used to establish nonlinear frequency band correlation mapping between electromagnetic distortion characteristic quantities and thermodynamic offset state quantities, calculate mutual information entropy values, and classify thermoelectric synergy imbalance levels based on the evolution of mutual information entropy values; Harmony generation module: used to generate a tuning instruction set with adaptive harmonic attenuation weights according to the thermal power coordination imbalance level; Magnetic field compensation module: used to analyze the transient disturbance of the ambient magnetic field and the full-band modulation distortion of the arc voltage. When the phase amplitude of the ambient magnetic field disturbance mismatches the attenuation gradient of the arc light intensity, a magnetic field compensation factor is generated. Calibration and adjustment module: used to reversely superimpose the magnetic field compensation factor into the tuning instruction set, and dynamically correct the wire feed rate calibration based on the spectral matching degree between the molten pool oscillation energy gradient and the harmonic distribution; Spectrum control output module: used to reconstruct the harmonic energy ratio spectrum of the tuning instruction set and output anti-interference welding control signals.

Images